@mayintoronto @mhoye @Rycaut @amydiehl
i just stumbled upon something where a researcher gave Claude a bunch of abstracts in genetics, to which the researcher had *randomly* assigned author names, and asked Claude to find patterns re research topics and author social identity. And Claude did. E.g., "Female authors" were more likely to produce research about "DEI, Ethics, and Social Dimensions of Genomics". It looks like there's no way to use these tools w/o *imposing cultural patterns*